Time and space bounds for multi-input machines #
The exact-run relation MultiInputTM.ComputesInTimeAndSpace mirrors the vendored CSLib
Turing.MultiTapeTM.ComputesInTimeAndSpace clause for clause (the machine has halted by
step t, has emitted exactly output, and has visited exactly s work-tape cells); the
equivalence with the one-input model is proved in MIPRE.TM.MultiInput.OneInputEquiv.
The packaged form MultiInputTM.ComputesFunWithBounds bounds time and space by arbitrary
functions of the tuple of input lengths. Deliberately, no convention such as "time as a
function of ∑ⱼ |xⱼ|" or "of maxⱼ |xⱼ|" is baked in at this level: the paper states
verifier runtimes primarily as functions of a single index n with separate caps on the
other inputs (blueprint def:lambda-bounded), and any such shape instantiates the length
tuple directly. See planning/tm-infrastructure.md, decision D4.
A proof that the Turing machine tm on input tuple input outputs output in at
most t steps and uses exactly s space. Neither the alphabet nor the state set is
required to be finite.
Equations
- One or more equations did not get rendered due to their size.
Instances For
A proof that the Turing machine tm computes the function f on i-tuples of
strings over an input/output alphabet embedded into the machine alphabet, within time
timeBound and space spaceBound — both arbitrary functions of the tuple of input
lengths.
Equations
- One or more equations did not get rendered due to their size.